RESEARCH ARTICLE

Lineage-specific proteins essential for endocytosis intrypanosomesPaul T. Manna1, Samson O. Obado2, Cordula Boehm1, Catarina Gadelha3, Andrej Sali4, Brian T. Chait2,Michael P. Rout2 and Mark C. Field1,*

ABSTRACTClathrin-mediated endocytosis (CME) is the most evolutionarilyancient endocytic mechanism known, and in many lineages thesole mechanism for internalisation. Significantly, in mammaliancells CME is responsible for the vast bulk of endocytic flux andhas likely undergone multiple adaptations to accommodate specificrequirements by individual species. In African trypanosomes, wepreviously demonstrated that CME is independent of the AP-2adaptor protein complex, that orthologues to many of the animal andfungal CME protein cohort are absent, and that a novel, trypanosome-restricted protein cohort interacts with clathrin and drives CME. Here,we used a novel cryomilling affinity isolation strategy to preservetransient low-affinity interactions, giving the most comprehensivetrypanosome clathrin interactome to date. We identified thetrypanosome AP-1 complex, Trypanosoma brucei (Tb)EpsinR,several endosomal SNAREs plus orthologues of SMAP and theAP-2 associated kinase AAK1 as interacting with clathrin. Novellineage-specific proteins were identified, which we designateTbCAP80 and TbCAP141. Their depletion produced extensivedefects in endocytosis and endomembrane system organisation,revealing a novel molecular pathway subtending an early-branchingand highly divergent form of CME, which is conserved and likelyfunctionally important across the kinetoplastid parasites.

KEY WORDS: Trypanosoma, Trafficking, Clathrin, Cytoskeleton,Flagellum, Evolution, Morphology

INTRODUCTIONEukaryotes originated from a lineage within the Thaumarchaeota,Aigarchaeota, Crenarchaeota and Korarchaeota (TACK) archaealclade (Guy and Ettema, 2011; Raymann et al., 2015). Subsequentelaboration of the eukaryotic cell during eukaryogenesis gave rise tothe nucleus, endomembrane system and acquisition of themitochondrion (Martin et al., 2015). Following emergence of atrue eukaryotic cell, the lineage rapidly diversified into multiplekingdoms or supergroups, represented for example, by plants,animals, fungi, amoeba and many protist lineages. The ∼1.5 billion

year period since this radiation is vast, and while core metabolic andgene expression pathways are frequently well conserved, manycellular features have experienced extensive specialisations, in partas a response to adaptive evolutionary forces.

One aspect of this diversity that has received considerable attentionis the endomembrane system, on account of this feature representingone of the more unique, yet flexible, aspects of eukaryotic cells. Theprimitive endomembrane system in the earliest eukaryotes gave riseto all of the endogenously derived internal compartments present inmodern lineages (Schlacht et al., 2014). As a consequence of thisextensive evolutionary history, compartments have experiencedsignificant divergence, resulting in vastly different morphologiesand functions, as reflected in diversification of the Golgi complex,lysosomes and endosomes. Furthermore, organelles unique tospecific lineages such as the rhoptries of Apicomplexa andglycosomes of kinetoplastids are now known and are derived fromendolysosomal ancestors and peroxisomes, respectively (Szöör et al.,2014; Tomavo, 2014). Residing within this compartmental variationare also apparently ancient and well-conserved molecular aspects oftrafficking systems, for example the exocytic apparatus comprisingthe exocyst and endocytic pathways mediated by clathrin (Field et al.,2007; Koumandou et al., 2007).

One of the more diverse eukaryotic phyla are the Kinetoplastida, agroup of flagellate protists, many of which are parasites. The Africantrypanosome Trypanosoma brucei lives and multiplies in thebloodstream and lymphatic system of its mammalian host, yetpersists in the face of challenge from both the innate and adaptiveimmune systems (Brun et al., 2010). Immune evasion relies onmultiple adaptations of the parasite cell surface composition andmembrane-trafficking dynamics (Manna et al., 2014). The canonicaldefence ofT. brucei is the expression of a dense coat of a single variantsurface glycoprotein (VSG) which, through antigenic variation (i.e.the switching of expression between VSG genes), periodically createsan antigenically distinct cell surface (Schwede et al., 2015). As anadditional defence, the VSG coat is constantly endocytosed andrecycled, removing surface-bound immune effectors (Engstler et al.,2007; Field and Carrington, 2009). Perturbations of endocyticrecycling severely limit parasite survival and infectivity in vitro andin vivo (Allen et al., 2003; Engstler et al., 2007; Natesan et al., 2011).

Common to all trypanosomatids is the polarisation of endocytosisand exocytosis to the flagellar pocket, a specialised plasmamembraneinvagination surrounding the base of the flagellum (Field andCarrington, 2009). This organelle is the gateway to and from the cellsurface and essentially controls the host–parasite interface, butsignificantly also contributes towards interactions with multipletherapeutics (Alsford et al., 2013; Zoltner et al., 2015). Less elaborate,but similar structures, are present at the base of the cilium in otherlineages, e.g. the proposed ciliary pocket in mammals and a ciliarypartitioning system in Tetrahymena (Benmerah, 2013; Ounjai et al.,2013). It is unknown whether all of these flagellum-associatedReceived 27 April 2016; Accepted 13 February 2017

1School of Life Sciences, University of Dundee, Dundee, Scotland DD1 5EH, UK.2The Rockefeller University, 1230 York Avenue, New York, NY 10021, USA. 3Schoolof Life Sciences, University of Nottingham, Nottingham NG2 7UH, UK. 4CaliforniaInstitute for Quantitative Biosciences, University of California, San Francisco, CA94158, USA.

*Author for correspondence (mfield@mac.com)

C.B., 0000-0002-1746-2450; M.C.F., 0000-0002-4817-6035

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1379

© 2017. Published by The Company of Biologists Ltd | Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

structures share common components and functionalities, and thepotential that the flagellar pocket contains parasite-specific featureslimits the utility of comparative approaches that facilitatedcharacterising much of the trypanosome endomembrane system(Field and Carrington, 2009). Recently the surface-exposed proteomeof the trypanosome cell surface, the flagellar pocket andendomembrane system have been described (Gadelha et al., 2015;Shimogawa et al., 2015). While the precise definition anddiscrimination between these interconnected membrane systems isunclear, what has emerged is a remarkable level of novelty within thetrypanosome surface membrane and endomembrane proteomes, withvery few surface proteins demonstrating significant conservationacross eukaryotes (Gadelha et al., 2015; Jackson et al., 2015). Thisextreme divergence potentially implies novel mechanisms of proteintrafficking.To directly explore the diversity of proteins associated with

flagellar pocket function, we previously described affinity isolation ofclathrin heavy chain [T. brucei (Tb)CHC] complexes usingpolyclonal antibodies, which identified several trypanosomatid-specific proteins that together mediating clathrin budding from theflagellar pocket membrane (Adung’a et al., 2013). However, owing tothe weak and transient nature of many clathrin–partner interactions,the analysis was most probably under-sampled, and many interactingproteinswere not isolated. Here, we applied a cryomillingmethod thatbetter preserves protein–protein interactions while physicallydisrupting the robust sub-pellicular microtubule corset of thetrypanosome cell (Obado et al., 2016a,b). We report an endocyticprotein cohort encompassing the majority of the expected earlyendocytic proteins, together with several novel kinetoplastid-specificgene products. Two of this latter group, which we designateTbCAP80 and TbCAP141, control not only clathrin-mediatedendocytosis (CME), but also the architecture and organisation ofthe broader endomembrane system. We suggest that these lineage-specific proteins are central to the coordination of membranetransport at the flagellar pocket, and are part of a growing list ofproteins that mediate the control of function in this crucial organelle.

RESULTSCryomilling andmass spectrometry identifies novel clathrin-interacting proteinsA previous analysis of clathrin-interacting proteins in trypanosomesyielded a series of lineage-specific proteins we termed clathrin-associated proteins or CAPs. Several CAPswere characterised in somedetail and demonstrated to mediate clathrin-budding events at theflagellar pocket membrane (Adung’a et al., 2013). However, manyexpected factors, including the adaptin AP-1 complex, were notretrieved despite being encoded in the trypanosome genome(Berriman et al., 2005). This suggests that additional proteins likelyinteract with clathrin, but were not captured by our earlier procedures,which used quite stringent conditions. Seeking a more complete viewof the clathrin interactome, we harvested, flash froze and cryogenicallylysed parasites by cryomilling them under liquid nitrogen to betterpreserve protein–protein interactions (Obado et al., 2016a,b). Insect-or bloodstream-form cells expressing CHC with a single allele taggedwith a C-terminal GFP (CHC::GFP) were subjected to cryomilling.CHC::GFP was affinity captured under a range of differing salt anddetergent conditions using polyclonal llama anti-GFP antibodiescoupled to magnetic beads (Fridy et al., 2014).Coomassie-stained SDS-PAGE of immunoprecipitates using

optimised conditions and 50 mg of cell powder (∼5×108cell equivalents) suggested candidate interacting proteins presentat substoichiometic levels, as expected (Fig. 1A). Mock

immunoprecipitation performed with beads to which no antibodyhad been coupled served as the non-specific binding control.Western blotting and matrix-assisted laser desorption/ionisationtime-of-flight mass spectrometry (MALDI-TOF MS) was used toverify the approach, and showed enrichment of known or predictedclathrin interactors (Adung’a et al., 2013; Borner et al., 2006, 2012,2014; Pearse, 1976) (Fig. 1A). Larger scale immunoprecipitationsusing 300 mg of starting material for each condition were analysedby liquid chromatography-electrospray ionisation massspectrometry (LC-ESI MS) and all identified proteins, from allconditions, are listed in Table S1. The top-ranked proteins, basedupon a probabilistic log(e) score for correct protein identificationcombined with label-free quantification [exponentially modifiedprotein abundance index (emPAI); Ishihama et al., 2005] from eachcondition were compared, and those proteins enriched in the anti-GFP antibody isolates versus the negative control selected (Table 1).

BLAST searches using the T. brucei strain 927 genome sequencedatabase as queries were used to identify orthologues for conservedproteins, which were verified by reciprocal BLAST. As well asmany known or predicted clathrin-interacting proteins, includingsubunits of the adaptor protein (AP)-1 complex and TbEpsinR(Allen et al., 2007; Gabernet-Castello et al., 2009; Tazeh et al.,2009), multiple partially characterised soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins(Murungi et al., 2014; Venkatesh et al., 2017) and proteins with noapparent homology to known human proteins were identified(Table 1).

Predicted domain architectures revealed several uncharacterisedsoluble cytosolic proteins, transmembrane-domain-containingSNARE proteins and the previously characterised trypanosomalGolgi-lysosomal glycoprotein 1 (TbGLP-1) (Fig. 1B) (Lingnauet al., 1999). In addition, TbEpsinR, Tb927.9.6560, Tb927.8.2820,Tb927.8.3840, Tb927.11.9180 and Tb927.10.13520 contain short,linear clathrin-binding motifs within predicted disordered regions(red dots in Fig. 1B), a common feature among many knownclathrin-interacting proteins (Kirchhausen et al., 2014).

In light of these analyses, five candidate proteins were selectedfor further investigation. As we were predominantly interested inpotential proteins that are cytoplasmic, soluble and associated withthe flagellar pocket, we chose only a single integral membraneprotein, TbGLP-1. The remaining four candidate proteins arepredicted to be cytosolic with folded N-terminal domains anddisordered C-termini containing predicted clathrin interactionmotifs. Tb927.9.6560 has an N-terminal serine/threonine kinasedomain and belongs to the same family of kinases as mammaliancyclin G-associated kinase (GAK) and AP-2 associated kinases(AAKs). These kinases are involved in clathrin-mediated traffickingprocesses through phosphorylation of the medium chains of theAP-1 and AP-2 adaptin complexes (Conner and Schmid, 2002;Ricotta et al., 2002; Umeda et al., 2000). This protein is hereafterreferred to as T. brucei AAK-like (TbAAKL). Tb927.11.9180(TbSMAP) has an N-terminal ArfGAP domain and belongs to theSMAP family of ArfGAPs implicated in clathrin-dependent vesicletransport (Tanabe et al., 2005; Schlacht et al., 2013). Of particularinterest were two uncharacterised proteins with no apparenthomology by BLAST to non-trypanosomatid proteins and noidentifiable domains, Tb927.8.2820 and Tb927.8.3840 (hereafterTbCAP140 and TbCAP80, respectively).

Localisation of putative clathrin-interacting proteinsFor immunolocalisation of candidate proteins in bloodstream-formtrypanosomes, the lifecycle stage with the highest endocytic activity

1380

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

and the clearest dependence upon endocytic function for viability,PCR-generated DNA fragments encoding triple haemagglutinintags were introduced into the endogenous genes via homologousrecombination. C-terminal tagging was verified by western blottingand immunofluorescence. All five candidate proteins colocalisedwith clathrin, although to varying degrees. This variablecolocalisation is to be expected given the diversity and functionalsegregation of inter-organelle membrane-trafficking events utilisingthe clathrin coat (Fig. 2A). An antibody to CHC widely labels theposterior region of the trypanosome cell, specifically between thetwo DAPI foci, the larger one representing the nuclear DNA and asecond smaller one representing the mitochondrial DNA, thekinetoplast. This clathrin-rich region of the cell contains thesingle Golgi complex and almost the entirety of the endocytic

compartments and, unsurprisingly, TbGLP-1 was observed here asmultiple punctae. The distribution of TbSMAP closely resembledthat of TbGLP-1, and both proteins colocalised extensively withCHC. The remaining three proteins were striking in their apparentlyrestricted codistribution with some of the clathrin-containingstructures. All three of TbAAKL, TbCAP80 and TbCAP141colocalised with a distinct subset of clathrin punctae located closeto the kinetoplast and therefore representing clathrin-coatedstructures at or near to the flagellar pocket membrane. Nascentclathrin-coated pits and vesicles are generally 100–200 nm indiameter, with several pits forming simultaneously at the flagellarpocket (Grünfelder et al., 2003). These clathrin structures are notreadily resolved by conventional light microscopy, making itchallenging to discern true associations with clathrin-coated

Fig. 1. Identification of T. brucei clathrin-associated proteins. Clathrin and itsassociated proteins were immunoprecipitatedfrom cryomilled powder generated from GFP-tagged CHC (CHC::GFP) bloodstream- orprocyclic-stage cells. (A) Coomassie-stainedSDS-PAGE shows bands specific to anti-GFPaffinity isolates versus controls, confirmed asknown CHC-associated proteins by westernblot and MALDI-TOF MS. (BSF, bloodstream-form parasites; PCF, procyclic-form parasites)(B) Domain organisation of selected high-confidence clathrin-associated proteinsidentified by LC-ESI MS (see Table 1 andTable S1). Domain predictions were generatedvia HHPRED, HMMER and InterProScan.

1381

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

membrane structures versus close apposition. We thereforeemployed structured illumination microscopy (SIM), whichdemonstrated a high level of colocalisation between all threecandidate proteins and clathrin-positive structures (Fig. 2B;Movies 1–3). This is in contrast to a known exocytic proteinTbSec15, which has been shown by SIM to form punctaeinterspersed with clathrin-positive structures at the flagellarpocket, but with essentially no colocalisation (Boehm et al., 2017).

Sequence and phylogenetic analysis of T. brucei clathrin-associated proteinsThe evolutionary distribution of candidate genes was investigated toassess their potential relevance to the broader trypanosomatidlineage. The phylogenetic distribution of the SMAP family ofArfGAPs has recently been characterised in detail and demonstratedto be pan-eukaryotic (Schlacht et al., 2013). Hence, we focused onthe remaining newly identified clathrin interactors, TbAAKL,TbCAP141 and TbCAP80. Orthologues of TbAAKL are foundacross the trypanosomatids and also in the sister group of free-livingBodo saltans, suggesting that the protein is conserved and anancient aspect of the kinetoplastid membrane-trafficking system(Fig. 3A). However, in trypanosomes that lack the AP-2 complexand also the closely related AP-2-retaining Tropidophorus grayi,TbAAKLmutations within the conserved D-(x)4-Nmotif of the VIbsubdomain are predicted to render the kinase inactive (Hanks andHunter, 1995) (Fig. 3B). This suggests a possible relaxation of therequirement for AP-2-directed kinase activity in this lineage and,perhaps most significantly, because this mutation seems to predatethe loss of the adaptor complex, it suggests the D-(x)4-N mutationcould enable subsequent loss of AP-2.

Because TbCAP80 and TbCAP141 were completelyuncharacterised, we first examined their predicted secondarystructures. Unexpectedly, these were strikingly similar despite noapparent sequence similarity, consisting of an extreme N-terminalthat is predominantly β-sheet containing, followed by a relativelylarge predominantly α-helical region containing putative coiled-coils (Fig. 3C). Downstream of the folded N-terminal domain, bothTbCAP80 and TbCAP141 are predicted to contain a largedisordered region. Domain prediction tools (HHpred, HMMERand InterproScan) returned no high-confidence predictions,although there is very weak support for the extreme N-terminal β-region representing a degenerate C2 domain. Furthermore, clathrin-binding motifs are located within the disordered regions of bothTbCAP80 and TbCAP141. Both are found as single copy genesacross the trypanosomatids and also in the sister group Bodonidae,but neither protein has identifiable orthologues outside of theDiscicristata (Fig. 3D). Significantly, this overall predictedarchitecture for TbCAP80 and TbCAP141 is reminiscent ofseveral characterised monomeric clathrin adaptors in which the N-terminal region is a membrane-binding domain while the C-terminus appears flexible and is predicted to be disordered;examples include the ENTH/ANTH domain epsin, epsinR andcalmodulin (CALM) family proteins, suggesting a generalarchitectural conservation consistent with clathrin binding(Kirchhausen et al., 2014). For functional assessment of CAP141and CAP80 domains, full-length and truncated forms were clonedinto vectors for expression in bloodstream-stage trypanosomes andmammalian cells (Fig. 4A) (Sunter et al., 2012). In bloodstream-form trypanosomes, full-length CAP141 and CAP80 GFP fusionslocalised to punctae at or near to the flagellar pocket, as was seen for

Table 1. High-confidence protein identifications from large-scale CHC cryo-immunopurification

T. brucei accession DesignationH. sapiensorthologue

Humanpredicted CCV

Molecularmass (kDa)

emPAI (PCF/BSF/Con) Reference(s)

Tb927.10.6050 TbCHC CHC Yes 190.6 6.4/3.6/1.0 Allen et al., 2003Tb927.10.14760 TbCLC CLC Yes 20.5 1.4/1.0/0.5 Gabernet-Castello et al., 2009Tb927.11.670 TbEpsinR EpsinR Yes 55.3 2.0/1.8/0.9 Gabernet-Castello et al., 2009Tb927.4.760 TbAP1γ AP1γ1 Yes 87.3 1.4/0.7/0.04 Allen et al., 2007Tb927.10.8040 TbAP1β AP1β1 Yes 76.3 1.2/0.2/NI Allen et al., 2007; Morgan et al., 2001Tb927.7.3180 TbAP1μ AP1μ1 Yes 49.2 1.8/1.2/0.05 Tazeh et al., 2009Tb927.9.6560 TbAAKL AAK Yes 80.2 1.6/0.5/0.4Tb927.8.2820 Hypothetical N/A N/A 141.1 2.1/0.4/0.03Tb927.8.3840 Hypothetical N/A N/A 80.1 1.9/0.3/NITb927.8.1870 TbGLP1 N/A N/A 67.6 2.6/1.9/1.1 Lingnau et al., 1999Tb927.11.9180 TbSMAP SMAP1/2 Yes 31.7 3.4/3.3/NITb927.10.6780 TbVPS45 VPS45 Yes 65 3.3/0.1/NITb927.10.1830 TbQc2,

TbSyntaxin6-likeSyntaxin6* Yes 25.8 3.3/0.7/NI Murungi et al., 2014; D.V. and M.C.F.,

unpublished observationTb927.5.3560 TbVAMP7b VAMP7* Yes 24.4 2.4/2.0/0.8 Murungi et al., 2014; D.V. and M.C.F.,

unpublished observationTb927.10.790 TbVAMP7c VAMP7* Yes 26.5 1.1/1.1/0.2 Murungi et al., 2014, D.V. and M.C.F.,

unpublished observationTb927.8.3470 TbQb2, TbVti-like Vti1* Yes 26.4 2.5/2.2/0.2 Murungi et al., 2014; D.V. and M.C.F.,

unpublished observationTb927.9.13030 TbSyn16b Syntaxin 16* Yes 36 2.3/0.29/NI Murungi et al., 2014, D.V. and M.C.F.,

unpublished observationTb927.10.5040 Hypothetical N/A N/A 30.9 5.5/5.5/NITb927.6.3940 TbATG27 N/A N/A 39.2 0.9/1.2/0.25Tb927.10.13520 Hypothetical N/A N/A 57.3 0.4/0.7/NI

Gene products identified as likely CHC interaction partners in T. brucei. Entries in bold are the focus of the current study.Where the encoded proteins have humanorthologues these have been indicated together with their status as predicted clathrin-coated vesicle (CCV)-associated proteins according to Borner et al. (2012,2014). Exponentially Modified Protein Abundance Index (emPAI) scores indicate enrichment in immunoprecipitations from CHC::GFP versus control [procyclicform (PCF), bloodstream form (BSF) and control (Con)]. NI denotes protein not identified in control immunoprecipitation; N/A indicates lack of human orthologuesand therefore no prediction; asterisks indicate homology assigned by BLAST rather than phylogeny. D.V., D. Venkatesh, School of Life Sciences, University ofDundee, UK.

1382

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

the endogenously tagged proteins. However, this flagellar pocketlocalisation was apparently readily saturable, likely reflecting alimited number of available binding sites, with higher expressionlevels leading to mis-localisation when compared to endogenouslyexpressed fusion proteins (Fig. 4B; Fig. S1). Removal of thepredicted disordered domains, containing the putative clathrin-binding motifs, from either protein had no apparent effect upon theflagellar pocket localisation, whereas expression of only the extremeN-terminal β-rich region as a GFP fusion led to a loss of flagellarpocket targeting (Fig. 4B).In a further attempt to determine whether the predicted

architectures of TbCAP80 and TbCAP141 functionallycorrespond to an N-terminal membrane-addressing domain andpotential recognition of specific lipid species at the membrane, weexpressed TbCAP80 and TbCAP141 as GFP fusion proteins inCOS7 cells (Fig. 4A). This approach was chosen as expression inthe mammalian cell background presumably isolates each proteinfrom trypanosome interaction partners that may influencelocalisation as well as representing a better-characterisedmembrane system, where the localisations of specific

phosphoinositide species are well documented. Full-lengthTbCAP141 had a largely perinuclear and reticular distribution(Fig. 4C), although with some clear plasma membrane association.The internal localisation for TbCAP141 may be a result ofoverexpression as seen in the homologous system. Thetrypanosome expression studies suggest that the large disorderedregion of this protein is likely dispensable for membranelocalisation. Therefore, a truncated version of TbCAP141,encoding only the N-terminal predicted folded domain (residues1–440) fused to GFP, was expressed. This construct showed a muchmore diffuse cytosolic signal while also retaining plasma membraneassociation (Fig. 4C). Finally, we expressed the extreme N-terminusof TbCAP141, corresponding to the potential C2-like fold (residues1–165), as a GFP fusion. This construct failed to localise to theflagellar pocket in bloodstream-form trypanosomes but showed themost pronounced plasmamembrane localisation, suggesting that theN-terminal β-rich domain contains a membrane localisation signal(Fig. 4C). We note, however, that in the absence of a fullunderstanding of the lipid environment at the trypanosomeflagellar pocket these data are merely supportive of this conclusion.

Fig. 2. Localisation of selected T. brucei clathrin-associated proteins. C-terminal triple HA tags were added to selected clathrin-associated proteins atendogenous genomic loci via homologous recombination in bloodstream-stage parasites. (A) Immunofluorescence colocalisation showed correlation betweenclathrin-associated proteins (red, left panels) and CHC (green, central panels). Note that TbAAKL (i), TbCAP80 (ii) and TbCAP141 (iii) showed preferentialcolocalisation with clathrin punctae in the vicinity of the flagellar pocket. DAPI staining is in blue. N, nucleus; K, kinetoplast (indicated on top panel only). Scale bar:2 μm. (B) Super-resolution SIM shows close association of clathrin-associated proteins (red, left panels, anti-HA antibody) with the flagellar pocket and clathrin-positive structures (green, centre panels, anti-TbCHC antibody) close to the kinetoplast. Boxes below the main panels show z-plane resections along the axisdenoted by the black arrowhead for each image. Scale bar: 500 nm.

1383

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

In contrast to TbCAP141, TbCAP80 constructs encoding bothfull-length protein and the extreme N-terminal β-rich domain fusedto GFP were largely cytosolic and lacked clear membraneassociation (Fig. 4C). This suggests that the targeting informationrequired for localisation of TbCAP141 to the plasma membrane, orthe flagellar pocket membrane in the native trypanosome cell, isconserved across eukaryotes, whereas TbCAP80 likely relies upon atrypanosome-specific factor as an accessory determinant.

TbSMAP, TbCAP80 and TbCAP141 are essential forendocytosisIdentification via CHC-GFP immunoprecipitation and colocalisationwith clathrin-containing structures suggest endocytic functions forTbAAKL, TbSMAP, TbCAP80 and TbCAP141.Many characterisedendocytic proteins in T. brucei are essential for viability inbloodstream-form trypanosomes, as is endocytosis itself,suggesting that if TbAAKL, TbSMAP, TbCAP80 and TbCAP141are indeed major players in endocytic trafficking, they should alsohave a significant impact on proliferation and viability.To address these questions, we generated inducible RNAi cell

lines in endogenous-locus-tagged bloodstream-form backgroundsfor each candidate protein. As for the localisation studies, we have

focused on bloodstream-form parasites as this life cycle stage hasthe highest endocytic activity as well as the clearest link betweenendocytic activity and viability. Following RNAi induction, proteindepletion was assayed by western blotting and cell proliferation wasrecorded (Fig. 5A). TbSMAP depletion led to a proliferativeinhibition after ∼24 h. Depletion of either TbCAP80 or TbCAP141produced strong proliferative inhibition from as little as 12 h postinduction, which is consistent with previous whole-genome studies(Alsford et al., 2011). By contrast, despite strong protein depletion,silencing of TbAAKL had no obvious effect upon parasiteproliferation. Furthermore, the proliferative defects followingTbSMAP, TbCAP80 or TbCAP141 depletion were accompaniedby morphological abnormalities (Fig. 5B). In particular, cellsbecame round and this was frequently accompanied by flagellarpocket swelling, with a large vacuolar structure resulting fromfailure of CME at the flagellar pocket seen under phase-contrastconditions, although this can also potentially arise from an impacton many trafficking events (Adung’a et al., 2013; Allen et al., 2003;Spitznagel et al., 2010). These morphological effects were absent inTbAAKL-depleted cells, but more severe and faster in onsetfollowing TbCAP80 and TbCAP141 depletion compared to upondepletion of TbSMAP.

Fig. 3. Sequence and phylogenetic analysis of newly identified clathrin-associated proteins. (A) Phylogenetic tree of the TbAAKL family across thekinetoplastids. Red arrowsmark potential occurrences of mutations predicted to render the kinase inactive. (B) Sequence alignment showing putative inactivatingmutations in kinetoplastid TbAAKL family kinases (asterisks). (C) Secondary structure predictions for TbCAP141 and TbCAP80. Note similarities with predictedextreme N-terminal β-rich regions followed by α-rich regions containing putative coiled coils and large disordered domains at the C-termini. (D) Phylogeneticanalyses of TbCAP141 and TbCAP80 families showing distribution across the kinetoplastids. Trees shown are best-scoring maximum likelihood trees (PhyML).Black branches are supported >80, >80 or >0.99 (as assessed by RaxML, PhyML abd MrBayes, respectively), grey branches are below this level of support.

1384

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

Fig. 4. Heterologous expression of TbCAP::GFP fusions. (A) C-terminal GFP fusionswere constructed for full-length TbCAP141 and TbCAP80, and truncatedproteins bearing the entire predicted N-terminal globular domain and the predicted β-rich extreme N-terminus alone. (B) Expression of TbCAP full-length andtruncation constructs in bloodstream-form parasites under the control of tetracycline. Low level expression necessitated amplification of native GFP fluorescenceby staining with anti-GFP followed by fluorescent secondary antibodies. Scale bar: 5 μm. (C) Expression of TbCAP full-length and truncation constructs in COS-7cells. Scale bar: 20 μm.

1385

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

As the enlarged flagellar pocket phenotype of TbSMAP-,TbCAP80- or TbCAP141-depleted cells suggested that there wasendocytic inhibition, we assayed this directly by following theuptake of fluorescently labelled probes. Transferrin is selectivelyacquired via the transferrin receptor (Steverding, 2000) whereas thelectin concanavalin A (ConA) is a non-selective binder of mannose-containing cell surface glyocproteins, predominantly VSG in thetrypanosome (Allen et al., 2003). When compared to non-inducedcontrol cells, the uptake of both ligands over a 45 min period wasstrongly inhibited, following depletion of TbSMAP, TbCAP80 andTbCAP141 (Fig. 5C,D). Consistent with this absence of aproliferative or morphological phenotype, TbAAKL depletion had

no apparent impact upon endocytic cargo uptake, indicating thatTbAAKL is not required for maintaining endocytic transport.

TbCAP141 and TbCAP80 are required for multiple aspectsof flagellar pocket functionThe effects of depleting several endocytic proteins are wellcharacterised at the ultrastructural level, and typically lead toenlargement of the flagellar pocket membrane area and luminalvolume, together with additional changes associated with deliveryor removal of membrane carriers to or from the cell surface(Adung’a et al., 2013; Allen et al., 2003; García-Salcedo et al.,2004; Manna et al., 2015).

Fig. 5. Phenotypic consequences of clathrin-associated protein depletion. Tetracycline-inducible RNAi cell lines for selected clathrin-associated proteinswere generated in HA-tagged bloodstream-form cells. (A) Effects of tetracycline addition on protein levels as assessed by western blotting before RNAi inductionand 24 h after induction using an anti-HA antibody (insets, - and +24, respectively), and its effect on cell proliferation expressed as mean±s.e.m. from threeindependent repeats. – (dashed lines), uninduced controls. (B) Morphology of induced RNAi cells at the indicated time points. Blue, DAPI staining; whitearrowheads denote vacuolar structures viewed under phase-contrast conditions that are reminiscent of swollen flagellar pockets. (C,D) Effects of clathrin-associated protein depletion on uptake of FITC–transferrin or FITC–ConA. (C) Quantification of FITC–transferrin or FITC–ConA uptake in induced RNAi cellsversus uninduced controls. Data are mean±s.d. of at least 50 cells per condition from two independent experiments normalised to non-induced controls.(D) Representative images showing FITC–transferrin or FITC–ConA (green) accumulation following a 45-min pulse. Blue, DAPI staining. Scale bars: 5 μm.*P<0.05, **P<0.005, ***P<0.001 versus control (t-test).

1386

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

We therefore prepared ultrathin sections of bloodstream-formcells depleted of TbSMAP, TbCAP80 and TbCAP141 by fastisothermal fixation for transmission electron microscopy (TEM).Control cells show the polarised distribution of the T. bruceiendocytic system, with the posterior flagellar pocket oftenassociated with clathrin-coated pits and vesicles (Fig. 6A). Thevacuoles that are seen under phase-contrast light, when observed by

conventional microscopy in TbSMAP-, TbCAP80- and TbCAP141-depleted cells (Fig. 5), can be confirmed to represent swelling of theflagellar pocket (Fig. 6A). While TbSMAP-depleted cells displayednormal endosome and Golgi complex profiles with clear tubularsorting endosomes present (Fig. 6B,C), extensive vacuolisation ofthe cytoplasm was seen in cells depleted of either TbCAP80 orTbCAP141 (Fig. 6B). These swollen vacuolar structures were

Fig. 6. Effects of clathrin-associated protein depletion on endomembrane system ultrastructure. Representative transmission electron micrographs ofultrathin resin sections of bloodstream-form parasites after RNAi-mediated depletion of clathrin-associated proteins. (A) Gross ultrastructural defects, in particularswelling of the flagellar pocket (FP) and cytoplasmic vacuolisation. (B) Tubular endosomes (black arrowheads) are apparent in the vicinity of the flagellar pocket ofcontrol and TbSMAP-depleted cells. A clathrin-coated pit (CCP) profile is seen at the flagellar pocket membrane of the control section (black arrowheads).Extensive vacuolisation (asterisks) is apparent in both TbCAP80- and TbCAP141-depleted cells. (C) Golgi (G) profiles. The control cell shows numerous ER–Golgi transport intermediates. TbSMAP depletion has little apparent effect upon Golgi ultrastructure, whereas depletion of TbCAP80 or TbCAP141 causesvacuolisation and swelling (asterisks) of the trans-cisternae. (D) Clathrin-coated profiles (white arrowheads) of control and TbCAP80- or TbCAP141-depletedcells. Aberrant large and flat coated profiles are seen in TbCAP80- or TbCAP141-depleted cells.

1387

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

prominent in the vicinity of the flagellar pocket and also associatedwith the trans face of the Golgi stack (Fig. 6C) indicating morewidespread endomembrane system defects. Finally, we examinedthe morphology of clathrin-coated structures associated with theflagellar pocket. Control cells displayed many clear clathrinstructures, ranging from early, shallow pits to deeply invaginatedand pre-scission necked pits (Fig. 6D). No such structures could beidentified in TbSMAP-depleted cells, whereas TbCAP80- andTbCAP141-depleted cells formed extended shallow pits. Thissuggests that the endocytic defects in these cells arise from a failureof curvature generation and coated pit progression. Depletion of thecharacterised clathrin adaptor proteins TbCALM and TbEpsinRproduces a similar phenotype (Manna et al., 2015), supportingcommon and non-redundant functions for these proteins. Althoughof uncertain significance, we also note that in TbCAP80- andTbCAP141-depleted cells, the enlarged flagellar pockets wereoccasionally (2 of 28 thin sections, 1 of 20 thin sections,respectively) associated with subpellicular microtubules, whichare a non-canonical feature of this specialised membrane domain,likely due to loss of differentiation from the plasma membrane (Fig.S2). Furthermore, depletion of either TbCAP80 or TbCAP141 alsoled to the appearance of cytoplasmic axonemes, often associatedwith a paraflagellar rod (3 of 28 and 5 of 20 thin sections,respectively) (Fig. S2). Although this observation could suggest aloss of cell polarity, and has been observed upon knockdown offunctionally related proteins (Price et al., 2007), this phenotype hasalso been observed to be a consequence of ablation of proteins withfunctions apparently unrelated to those studied here (Morgan et al.,2005; Baines and Gull, 2008; Morriswood and Schmidt, 2015;Sheader et al., 2005). In summary, the most representativeultrastructural defects observed suggest crucial roles for TbCAP80and TbCAP141 in both endocytosis and Golgi complex andflagellar pocket function and are notable by their extensive scope,impacting a considerable proportion of the endomembrane system.

DISCUSSIONTrypanosomes likely diverged over one billion years ago and manyexamples of extreme specialisations within membrane trafficking,gene expression and nuclear organisation in these organisms areknown (e.g. Daniels et al., 2010; DuBois et al., 2012; Field andCarrington, 2009; Obado et al., 2016b). Comparative genomicsdemonstrates that trypanosome endocytosis is divergent (Adung’aet al., 2013; Field et al., 2007; Manna et al., 2013, 2015), based bothon the absence of orthologues of multiple endocytic genes that arepresent in animals and fungi, and recent identification of severallineage-specific proteins that are associated with clathrin. We havefurther investigated the clathrin interactome as our earlier studyfailed to identify many expected and conserved gene products. Byperforming cryomilling and improved affinity purification, weindeed identified many of these anticipated interaction partners(Fig. 7) and more lineage-specific proteins, expanding the evidencethat novel mechanisms are operating in these cells.

Specifically, we recovered all subunits of AP-1, many SNAREsexpected to be associated with the endosomal/post-Golgi traffickingsystems, trypanosome orthologues of AAK and SMAP, as well asHsc70 and TbEpsinR, with these latter two also being identified inour earlier work (Adung’a et al., 2013). We also recovered Vps45,the endosomal Sec1/SM protein, which in higher eukaryotesinteracts with multiple SNAREs and which, therefore, may wellhave acted as a bridge between clathrin and the SNARE proteinsrecovered in this screen. Vps45 also associates with Rabenosyn5 inmammalian cells, providing an additional connection to earlyendosomes. The SNAREs in particular are likely associated withendosomal and recycling trafficking based on recent phylogeneticsand immunoisolation investigations (Venkatesh et al., 2017). Wealso recovered TbGLP-1, an abundant trypanosome-specifictransmembrane domain protein associated with the Golgi complex(Lingnau et al., 1999). The identification of several transmembranedomain proteins, together with AP-1 and Vps45, gives us

Fig. 7. Interactome subtending CME in trypanosomes. Whereas clathrin is distributed widely throughout the trypanosome endomembrane system, flagellarpocket-specific factors act in the earliest stages of endocytosis. TbCALM and TbEpsinR are likely recruited via localised lipid accumulation. TbCAP141 andTbCAP80 are also enriched in flagellar pocket-associated clathrin structures. We propose that the large disordered regions of these proteins likely contribute tothe generation of local membrane curvature through molecular crowding. Additionally, a conserved core of clathrin-associated accessory proteins (TbVps45,TbAAKL and TbHsc70) likely act to coordinate clathrin coat dynamics and regulate the sorting of the conserved SNARE machinery.

1388

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

confidence that we have now sampled the trypanosome endosomalmachinery to a considerable depth, as does the observation that avery similar cohort of CME proteins are present in T. cruzi (Kalbet al., 2016). It is of interest that we detect AP-1 and not any otheradaptin complex proteins, and it is likely that AP-1 is involved inpathways delivering multiple proteins to the lysosome, but notdirectly in endocytosis from the plasma membrane (Allen et al.,2007; Morgan et al., 2001; Zoltner et al., 2015). The trypanosomeorthologue of the autophagy protein Atg27 is also of interest; thisprotein lacks a mammalian homologue but has been characterised inyeast where it cycles between the Golgi and peripheral membranestructures, and controls trafficking of another autophagy proteinAtg9 (Backues et al., 2015; Yen et al., 2007).The current study, and recent publications from our laboratory

and others, allow us to assemble a model for a protein interactionnetwork underlying flagellar pocket endocytic function (Fig. 7)(Adung’a et al., 2013; Demmel et al., 2014; Gabernet-Castello et al.,2009; Manna et al., 2015). We suggest that the local generation of aphosphoinositide signal, likely phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2], is an essential step in clathrin-coated pitformation, and this signal is sensed by the phosphoinositide-bindingclathrin adaptors TbEpsinR and TbCALM, and possibly TbCAP80and TbCAP141 (Demmel et al., 2014; Manna and Field, 2016;Manna et al., 2015). T. brucei homologues of Hsc70 and GAK/AAK family kinases are also associated with clathrin structures buttheir roles remain unclear due to apparently minor phenotypesfollowing RNAi-mediated depletion.TbCAP80 and TbCAP141 have essential roles in CME; that these

apparently abundant proteins were not identified in our earlier studyhighlights the improved capacity for identifying likely weak andtransient interactions that cannot be seen in approaches based oncryogenic cell disruption. Depletion of either TbCAP80 orTbCAP141 gives strikingly similar phenotypes, suggesting relatedyet non-redundant functions. This is further supported by similarityin predicted secondary structure. Whilst unrelated at the sequencelevel to canonical monomeric clathrin adaptors, such as epsin orCALM/AP180, TbCAP80 and TbCAP141 do share a strikingsimilarity of domain organisation, with folded N-terminal domainsand disordered C-terminal regions that harbour predicted clathrin-binding motifs. This is suggestive of a degree of mechanisticsimilarity, and we have shown that the N-terminal folded domain ofTbCAP141 carries a membrane-targeting signal which mayconceivably be functionally analogous to the ENTH and ANTHdomains of epsin and CALM proteins. Further support for thisproposition is found in the similarity of specific defects seen inclathrin-coated pit morphology following depletion of TbCAP80,TbCAP141, TbCALM or TbEpsinR. It appears that ablation of anyof these proteins leads to endocytic inhibition through a block ofendocytic pit maturation, causing an accumulation of broad, flatclathrin-coated structures at the flagellar pocket. Interestingly, bothwhole-genome RNAi and proteomics studies indicate thatTbCAP80 and TbCAP141 are also essential in the procyclic form,but are expressed at somewhat lower levels (∼0.36-fold and ∼0.25-fold, respectively, relative to in the bloodstream form), suggestingthat they also have roles in endocytic transport in the insect form(Alsford et al., 2011; Urbaniak et al., 2012).That depletion of any of the proteins that are suggested to

independently drive clathrin recruitment causes malformed pits,rather than a failure of clathrin assembly per se, may be explainedby an emerging model of membrane vesicle formation. In additionto proposed membrane-curvature-driving domains, such asamphipathic helices, the steric interactions between coat proteins

themselves are suggested to generate positive curvature throughmolecular crowding (Busch et al., 2015; Copic et al., 2012; Milleret al., 2015; Stachowiak et al., 2012). The concentration of luminalcargoes packaged into forming pits generates an opposing force thatmust be overcome during coat assembly. T. brucei, with its denseVSG coat, would therefore be predicted to be particularly sensitiveto perturbations of this curvature-driving force provided bymolecular crowding of coat proteins. We suggest therefore thatdepletion of any one of the major components of the endocytic coatin T. brucei, while likely not sufficient to prevent clathrin assembly,leads to a failure of pit closure and subsequent endocytic inhibition.

Despite being heavily studied in yeast and metazoan systems,attention has only relatively recently turned to the degree of lineage-specific adaptation in the molecular mechanisms of CME. Incontrast to the adaptin-related TPLATE/TSET complex recentlycharacterised in plants and Amoebozoa (Gadeyne et al., 2014; Hirstet al., 2014), TbCAP80 and TbCAP141 appear to be true lineage-specific innovations. This adds to the remarkably large number oflineage-specific clathrin-associated proteins with roles intrypanosome endocytosis (Adung’a et al., 2013) and is furtherevidence for significant remodelling of endosomal systems toenable specific adaptation, in this instance facilitating parasitism.While less extensive than the evolution of rhoptries and densegranules as unique organelles intimately associated with cellinvasion in Apicomplexa (Tomavo, 2014), it is clear that theendosomal system of trypanosomes plays vital roles in thepathogenesis of these organisms.

In summary, this study reveals the molecular pathwayssubtending an evolutionarily early branching and highly divergentform of CME that arose via the acquisition of a largely novel set ofgene products including TbCAP80 and TbCAP141, which areconserved and likely functionally important across the kinetoplastidparasites.

MATERIALS AND METHODSCell culture and transfectionProcyclic-form (PCF) Lister 427 Trypanosoma brucei brucei, bloodstream-form (BSF) Lister 427 and single-marker bloodstream-form (SMB)parasites were cultured as previously described (Brun and Schönenberger,1979; Hirumi and Hirumi, 1989; Wirtz et al., 1999). For generation andmaintenance of lines harbouring selectable markers, antibiotics were used atthe following concentrations: G418, 1 μg/ml; puromycin, 0.2 μg/ml; andhygromycin B, 5 μg/ml. COS-7 cells were cultured in Dulbecco’s modifiedEagle’s medium (DMEM) with 10% fetal bovine serum, 100 U/mlpenicillin, 100 U/ml streptomycin and 2 mM L-glutamine. Fortransfection of bloodstream-stage parasites, 3×107–4×107 cells weretransfected with 10 μg of DNA using an AMAXA nucleofection systemand the human T-cell nucleofection kit (Lonza). Monoclonal populationswere obtained by limiting dilution. Procyclic-stage parasites weretransfected by electroporation with a Bio-Rad Gene Pulser (1.5 kV, 25μF). COS-7 cells were transfected with Fugene HD according tomanufacturer’s instructions.

RNAi and endogenous-locus taggingProcedures were carried out as described previously (Alibu et al., 2005;Manna et al., 2015; Oberholzer et al., 2006). Endogenous locus taggingcassettes were generated by PCR using the pMOTag system and gene-specificprimer pairs as follows: TbCHC F, 5′-CCACCCGCAACCGGGCTACGG-TGGTGTGCCCGGTCAGGGATATGCTGGAGGGATGGGAAACCCT-AACATGATGCCATACGGTACCGGGCCCCCCCTCGAG-3′, TbCHCR, 5′-ATTTCTTCCCCTCCACCTACTCACCCTTTTCTCCTCCCATCT-CCCTTCCCTGTGTTTCTTTTGTCCTTTTGGGCTGGCGGCCGCTCT-AGAACTAGTGGAT-3′; TbCAP141 F, 5′-GAGTTCGCTCATCCTCCT-GGTCGTGGTGTATGTGCTGAATGAGGGTCGTTATACACCTTTTG-

1389

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

CCGCCCGTTTTCCAGTAGGTACCGGGCCCCCCCTCGAG-3′, TbC-AP141 R, 5′-CACCAAACCAGCCTTGATATTAATTCCTTCACTTCC-TCTATACGGGTTCTTCATATCACTTCCACGAATGCAAGTGGCGG-CCGCTCTAGAACTAGTGGAT-3′; TbCAP80 F, 5′-CTTTGGGTCCGA-TCACCTCAGTGTCAGCAAAGATAAGAGGGAGAGTGGGAATCAC-ACGCTTACTTTCAACTTTGGTAGTGGTACCGGGCCCCCCCTCGA-G-3′, TbCAP80 R, 5′-AGCACATTGTTGCAGGCGTTAGAACCACTT-TTTATCTCTTTCCCTTGTGTGTTTTCACTATTCTTGAAGATACCT-GGCGGCCGCTCTAGAACTAGTGGAT-3′; TbSMAP F, 5′-TACTC-CGAGTAATCAAGGTCCTCCGCACGTATATAGTGCTTGGGCCCCA-TCCGGTTCCTCCAAATGCTTTTCTCCTCAGGGTACCGGGCCCCCC-CTCGAG-3′; TbSMAP R, 5′-ATTTCAGTAACTTTTTTTTGTAATTTT-GCAACTTCAACTCTTTCTTCGCTAATGTAGGGCCTTCACCGCGAG-TGTGGCGGCCGCTCTAGAACTAGTGGAT-3′; TbGLP1 F, 5′-TACAT-CTCCAACATCCCAACCGCCGGTGCGGCGGTGAACTCGGCCGGTA-CGAAAGGCTCCGTCATTGAGGTGGAGGATGGTACCGGGCCCC-CCCTCGAG-3′, TbGLP1 R, 5′-TTTTCCTCGTTCCGGAAGGGTCCA-TGTTCTTCAAACACCCACTGCTACCTGGTACTCTCGTCAACCGTT-CAATGGCGGCCGCTCTAGAACTAGTGGAT-3′; and TbAAK F, 5′-G-GAAACGACCTATTCCAAAGGCCACAGCAGCAACAGCCACAGCA-ACCAGAGAAGGACCCCTTCGCCAGTCTCTTCAAGGGTACCGGG-CCCCCCCTCGAG-3′, TbAAK R, 5′-GACCCCCCCTTTTTTTTTGG-GGGGGGGGTGGCGTATGATGCTGTTCTGTGCCAGTATTGCAGT-CATGTAAAACATGGCGGCCGCTCTAGAACTAGTGGAT-3′. Correctintegration of the tagging cassette was verified by western blot. Gene-specific RNAi fragments were selected with RNAit (Redmond et al., 2014)and amplified from genomic DNA using the following primer pairs:TbAAK F, 5′-TTCTGCTTCTCGCAGACTGA-3′, TbAAK R, AAGAG-GTCATCCGTTGTTGG-3′; TbCAP141 F, 5′-GCAGTTGGAGGAGCT-ACTGG-3′, TbCAP141 R, 5′-TTTCTCTTTCGAAGTGCGGT-3′;TbCAP80 F, 5′-CATGCGGAAAGAAAACCAAT-3′; TbCAP80 R, 5′-G-CTCTTGTTTCTGTGGAGCC-3′; and TbSMAP F, 5′-CGAGGACGCA-GAAAAAGAAC-3′, TbSMAP R, 5′-TGGGCAAGTACTAACCTCGG-3′.PCR fragments were cloned into p2T7TAblue and constructs were verified bySanger sequencing. Multiple clonal RNAi lines were generated for eachconstruct in the corresponding genomic-locus-tagged SMB background andassayed for robust and reproducible protein loss following tetracycline(1 μg/ml) addition by western blotting prior to phenotypic characterisation.

Cryoimmunoisolation and proteomicsApproximately 1010 mid-log phase BSF or 5×1010 PCF parasites wereharvested by centrifugation (800 g for 15 min) and washed with chilled PBSsupplemented with 5.5 mM glucose. Cells were re-suspended in chilled PBSplus glucose with protease inhibitors and 10 mM dithiothrietol to a density of5×109 cells/ml and flash frozen in liquid nitrogen to preserve protein–proteininteractions as close to native state as possible. Cells were cryomilled into afinely ground powder in a planetary ball mill (Retsch). For a detailed protocolrefer to Obado et al. (2016a) or the National Center for Dynamic InteractomeResearch website (www.ncdir.org/public-resources/protocols/). Aliquots ofcryo-lysate powder were re-suspended in immunoprecipitation buffer[100 mM sodium phosphate pH 7.4, 250 mM Na-citrate, 0.5% Tween 20and protease inhibitors without EDTA (Roche)]. Subsequent affinity capturewas performed as previously described using either magnetic beads coupledto anti-GFP antibody or uncoupled beads to serve as a control for non-specificinteraction with the beads (Obado et al., 2016a,b). Affinity-purified proteinswere elutedwith 2%SDS in 40 mMTris-HCl pH8.0, reduced in 50 mMDTTand alkylated with 100 mM iodoacteamide prior to downstream analysis(SDS-PAGE followed by protein identification by using ESI orMALDI-TOFmass spectrometry). Eluates were fractionated on pre-cast Novex 4-12% BisTris gels (Life Technology), stained using colloidal Coomassie (GelCodeBlue–Thermo) and analysed by mass spectrometry as previously described(DeGrasse et al., 2009).

Immunofluorescence localisation and endocytosis assaysParasites were cultured, harvested and prepared as described previously(Manna et al., 2015). Anti-HA (3F10, Roche) was used at 1:1000 andpolyclonal rabbit anti-TbCHC (Morgan et al., 2001) was used at 1:2500. Forlocalisation of GFP-tagged TbCAP80 and TbCAP141 truncations in

bloodstream-form trypanosomes, the high-level protein expressionrequired to image native GFP fluorescence led to mislocalisation ascompared to endogenously expressed tagged proteins. Therefore, in theseexperiments, expression levels were restricted to near endogenous levels andthe fluorescence signal was boosted by staining with polyclonal rabbit anti-GFP (1:250, ab6556; Abcam) followed by Alexa-Fluor-488-conjugatedsecondary antibody (Invitrogen). Wide-field fluorescence images wereacquired using a Nikon Eclipse E600 microscope with a Hamamatsu ORCACCD camera and MetaMorph software (Universal Imaging, Marlow, UK).For structured illumination microscopy (SIM), images were acquired undera 60×, 1.42 NA objective using a DeltaVision OMX V3 system (AppliedPrecision, Preston, UK) and deconvolved using softWoRx 5.0 software(Applied Precision). Endocytic activity was assessed by fluorescent liganduptake over a 45 min period as described previously (Gabernet-Castelloet al., 2009; Manna et al., 2015). Quantification of fluorescence intensitieswas carried out in Fiji software (National Institutes of Health, USA, http://fiji.sc/Fiji). Image panels were prepared in Photoshop (Adobe).

Homology searches, domain predictions and phylogeneticanalysesFor the identification of homologous sequences and reconstruction ofprotein phylogenies, relevant excavate sequences were first identified. Toidentify related proteins from the kinetoplastids, BLAST searches werecarried out at TriTrypDB (http://tritrypdb.org) against Trypanosoma vivax,Trypanosoma cruzi, Trypanosoma brucei, Trypanosoma grayi,Trypanosoma congolense, Leishmania seymouri, Leishmania tarentolae,Leishmania mexicana, Leishmania major, Leishmania donovani,Leishmania infantum, Crithidia fasciculate and Endotrypanummonterogeii. Bodo saltans was searched at GeneDB (http://www.genedb.org), Naegleria gruberi was searched at the Joint Genome Institute (http://genome.jgi.doe.gov). Any identified excavate sequences were then used inan attempt to identify homologues from other eukaryotic supergroups.Initial protein-BLAST and protein-PSI-BLAST searches were carried out atNCBI (http://www.ncbi.nlm.nih.gov) against the non-redundant proteinsequence database or species-specific protein databases where appropriate.Where no clear homologues could be identified outside of the excavates,multiple protein alignments of excavate sequences were generated and usedas queries against the UniProt KB database via the Jackhmmer iterativesearch algorithm (http://www.ebi.ac.uk/Tools/hmmer/search/jackhmmer).Finally, all identified sequences were verified by reciprocal BLAST and/orjackhammer searches and manual inspection of alignments. To identifyconserved domains, retrieved sequences were run through HHPRED (http://toolkit.tuebingen.mpg.de/hhpred), hmmscan (http://www.ebi.ac.uk/Tools/hmmer/search/hmmscan) and InterProScan (http://www.ebi.ac.uk/interpro/search/sequence-search). For phylogenetic reconstruction, sequences werealigned using MergeAlign (Collingridge and Kelly, 2012) and edited forgaps and low conservation. The correct evolutionary model was assessedusing ProtTest (http://darwin.uvigo.es/software/prottest2_server.html) andphylogenetic trees were constructed using Bayesian (MrBayes, https://www.phylo.org) and maximum likelihood (PhyML, http://www.atgc-montpellier.fr/phyml/) approaches. MrBayes analyses were run using a mixed model forat least 106 generations and until convergence, assessed by standarddeviation of splits frequencies of <0.05.

Mammalian expression of TbCAP141 and TbCAP80 truncationsFor mammalian expression of full-length and truncated forms of TbCAP141and TbCAP80, the coding sequences were amplified from bloodstream-formLister 427 Trypanosoma brucei brucei genomic DNA and inserted into thepEGFP-N2 (Clontech) vector for C-terminal GFP tagging and mammalianexpression. TbCAP141 full-length was amplified using primers: forward,5′-GAGCTCATGCCGCTATACAATATCACTATTCA-3′ and reverse, 5′-GAATTCTACTGGAAAACGGGCGGC-3′. The N-terminal folded domain(residues 1–440)was amplified using the same forward primer and the reverseprimer 5′-GAATTCGGGACCGCCGGAGGCTGCT-3′. The extreme N-terminal putative β-rich region (residues 1–165) was cloned using the sameforward primer and the reverse primer 5′-GAATTCGAGGCGA-ACTGTTTCCACTG-3′. These PCR fragments were inserted between theSacI and EcoRI sites of pEGFPN-N2. TbCAP80 fragments were generated

1390

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

using the common forward primer 5′-AGATCTATGCCCGAGGC-GCATATT-3′ and reverse primers 1–421, 5′-AAGCTTGAACGGG-CCCATTTGAGCAC-3′ and 1–175, 5′-AAGCTTGAATCGGTAGCCAT-CCGCTT-3′. These fragments were cloned between BglII and HinDIII sites.For doxycycline-inducible expression in bloodstream-form trypanosomes atnear-endogenous levels, the corresponding constructs were cloned into a T7RNA polymerase-independent inducible expression plasmid as previouslydescribed (Sunter et al., 2012).

Thin-section transmission electron microscopySamples were prepared as previously described (Gadelha et al., 2009).Briefly, cells were fixed isothermally in culture with 2.5% (v/v)glutaraldehyde at 37°C prior to harvesting by centrifugation (800 g for 15min). Following fixation, samples were post fixed in 1% (w/v) osmiumtetroxide in PBS for 30 min at room temperature and en bloc stained with1% (w/v) aqueous uranyl acetate. Following dehydration through an acetoneseries, samples were embedded in epoxy resin. Ultrathin (70 nm) sectionswere post stained with 2% uranyl acetate and lead citrate. Samples wereimaged on a Tecnai G2 transmission electron microscope (FEI, USA).

AcknowledgementsWe thank Divya Venketesh for annotation of trypanosome SNAREs.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsM.C.F., P.T.M. conceived the study; P.T.M., C.B., S.O.O., C.G. performed theresearch and P.T.M., C.B., S.O.O., C.G., M.P.R. and B.T.C. analysed the data.P.T.M., S.O.O., C.G., M.P.R. and M.C.F. wrote and edited the manuscript.

FundingThis work was supported by the Wellcome Trust (grant 090007/Z/09/Z), MedicalResearch Council (grant G0900255 to C.G. and M.C.F.) and the National Institutesof Health (P41 GM109824 and R21 AI096069 to M.P.R.). Deposited in PMC forimmediate release.

Supplementary informationSupplementary information available online athttp://jcs.biologists.org/lookup/doi/10.1242/jcs.191478.supplemental

ReferencesAdung’a, V. O., Gadelha, C. and Field, M. C. (2013). Proteomic analysis of clathrininteractions in trypanosomes reveals dynamic evolution of endocytosis. Traffic 14,440-457.

Alibu, V. P., Storm, L., Haile, S., Clayton, C. and Horn, D. (2005). A doublyinducible system for RNA interference and rapid RNAi plasmid construction inTrypanosoma brucei. Mol. Biochem. Parasitol. 139, 75-82.

Allen, C. L., Goulding, D. and Field, M. C. (2003). Clathrin-mediated endocytosis isessential in Trypanosoma brucei. EMBO J. 22, 4991-5002.

Allen, C. L., Liao, D., Chung, W.-L. and Field, M. C. (2007). Dileucine signal-dependent and AP-1-independent targeting of a lysosomal glycoprotein inTrypanosoma brucei. Mol. Biochem. Parasitol. 156, 175-190.

Alsford, S., Turner, D. J., Obado, S. O., Sanchez-Flores, A., Glover, L.,Berriman, M., Hertz-Fowler, C. and Horn, D. (2011). High-throughputphenotyping using parallel sequencing of RNA interference targets in theAfrican trypanosome. Genome Res. 21, 915-924.

Alsford, S., Field, M. C. and Horn, D. (2013). Receptor-mediated endocytosis fordrug delivery in African trypanosomes: fulfilling Paul Ehrlich’s vision ofchemotherapy. Trends Parasitol. 29, 207-212.

Backues, S. K., Orban, D. P., Bernard, A., Singh, K., Cao, Y. and Klionsky, D. J.(2015). Atg23 and Atg27 act at the early stages of Atg9 trafficking in S. cerevisiae.Traffic 16, 172-190.

Baines, A. and Gull, K. (2008). WCB is a C2 domain protein defining the plasmamembrane - sub-pellicular microtubule corset of kinetoplastid parasites. Protist159, 115-125.

Benmerah, A. (2013). The ciliary pocket. Curr. Opin. Cell Biol. 25, 78-84.Berriman, M., Ghedin, E., Hertz-Fowler, C., Blandin, G., Renauld, H.,Bartholomeu, D. C., Lennard, N. J., Caler, E., Hamlin, N. E., Haas, B. et al.(2005). The genome of the african Trypanosome Trypanosoma brucei. Science309, 416-422.

Boehm, C., Obado, S., Manna, P. T., Gadelha, C., O’Reilly, A. J., Munson, M.,Chait, B. T., Rout, M. P. and Field, M. C. (2017). The trypanosome exocyst: a

conserved structure revealing a new role in endocytosis. PLoS Pathog. 13,e1006063.

Borner, G. H. H., Harbour, M., Hester, S., Lilley, K. S. and Robinson, M. S.(2006). Comparative proteomics of clathrin-coated vesicles. J. Cell Biol. 175,571-578.

Borner, G. H. H., Antrobus, R., Hirst, J., Bhumbra, G. S., Kozik, P., Jackson,L. P., Sahlender, D. A. and Robinson, M. S. (2012). Multivariate proteomicprofiling identifies novel accessory proteins of coated vesicles. J. Cell Biol. 197,141-160.

Borner, G. H. H., Hein, M. Y., Hirst, J., Edgar, J. R., Mann,M. andRobinson,M. S.(2014). Fractionation profiling: a fast and versatile approach for mapping vesicleproteomes and protein-protein interactions. Mol. Biol. Cell 25, 3178-3194.

Brun, R. and Schonenberger, (1979). Cultivation and in vitro cloning or procyclicculture forms of Trypanosoma brucei in a semi-defined medium. Shortcommunication. Acta Trop. 36, 289-292.

Brun, R., Blum, J., Chappuis, F. and Burri, C. (2010). Human Africantrypanosomiasis. Lancet 375, 148-159.

Busch, D. J., Houser, J. R., Hayden, C. C., Sherman, M. B., Lafer, E. M. andStachowiak, J. C. (2015). Intrinsically disordered proteins drive membranecurvature. Nat. Commun. 6, 7875.

Collingridge, P. W. and Kelly, S. (2012). MergeAlign: improving multiple sequencealignment performance by dynamic reconstruction of consensus multiplesequence alignments. BMC Bioinformatics 13, 117.

Conner, S. D. and Schmid, S. L. (2002). Identification of an adaptor-associatedkinase, AAK1, as a regulator of clathrin-mediated endocytosis. J. Cell Biol. 156,921-929.

Copic, A., Latham, C. F., Horlbeck, M. A., D’Arcangelo, J. G. and Miller, E. A.(2012). ER cargo properties specify a requirement for COPII coat rigidity mediatedby Sec13p. Science 335, 1359-1362.

Daniels, J.-P., Gull, K. and Wickstead, B. (2010). Cell biology of the trypanosomegenome. Microbiol. Mol. Biol. Rev. 74, 552-569.

DeGrasse, J. A., DuBois, K. N., Devos, D., Siegel, T. N., Sali, A., Field, M. C.,Rout, M. P. and Chait, B. T. (2009). Evidence for a shared nuclear pore complexarchitecture that is conserved from the last common eukaryotic ancestor. Mol.Cell. Proteomics 8, 2119-2130.

Demmel, L., Schmidt, K., Lucast, L., Havlicek, K., Zankel, A., Koestler, T.,Reithofer, V., de Camilli, P. and Warren, G. (2014). The endocytic activity of theflagellar pocket in Trypanosoma brucei is regulated by an adjacentphosphatidylinositol phosphate kinase. J. Cell Sci. 127, 2351-2364.

DuBois, K. N., Alsford, S., Holden, J. M., Buisson, J., Swiderski, M., Bart, J.-M.,Ratushny, A. V., Wan, Y., Bastin, P., Barry, J. D. et al. (2012). NUP-1 Is a largecoiled-coil nucleoskeletal protein in trypanosomes with lamin-like functions. PLoSBiol. 10, e1001287.

Engstler, M., Pfohl, T., Herminghaus, S., Boshart, M., Wiegertjes, G.,Heddergott, N. and Overath, P. (2007). Hydrodynamic flow-mediated proteinsorting on the cell surface of Trypanosomes. Cell 131, 505-515.

Field, M. C. and Carrington, M. (2009). The trypanosome flagellar pocket. Nat.Rev. Micro. 7, 775-786.

Field, M. C., Gabernet-Castello, C. and Dacks, J. B. (2007). Reconstructing theEvolution of the Endocytic System: Insights from Genomics and Molecular CellBiologyEukaryotic Membranes and Cytoskeleton. pp. 84-96. New York: Springer.

Fridy, P. C., Li, Y., Keegan, S., Thompson, M. K., Nudelman, I., Scheid, J. F.,Oeffinger, M., Nussenzweig, M. C., Fenyo, D., Chait, B. T. et al. (2014). A robustpipeline for rapid production of versatile nanobody repertoires. Nat. Methods 11,1253-1260.

Gabernet-Castello, C., Dacks, J. B. and Field, M. C. (2009). The single ENTH-domain protein of trypanosomes; endocytic functions and evolutionaryrelationship with epsin. Traffic 10, 894-911.

Gadelha, C., Rothery, S., Morphew, M., McIntosh, J. R., Severs, N. J. and Gull,K. (2009). Membrane domains and flagellar pocket boundaries are influenced bythe cytoskeleton in African trypanosomes. Proc. Natl. Acad. Sci. USA 106,17425-17430.

Gadelha, C., Zhang, W., Chamberlain, J. W., Chait, B. T., Wickstead, B. andField, M. C. (2015). Architecture of a host-parasite interface: complex targetingmechanisms revealed through proteomics. Mol. Cell. Proteomics 14, 1911-1926.

Gadeyne, A., Sanchez-Rodrıguez, C., Vanneste, S., Di Rubbo, S., Zauber, H.,Vanneste, K., Van Leene, J., De Winne, N., Eeckhout, D., Persiau, G. et al.(2014). The TPLATE adaptor complex drives clathrin-mediated endocytosis inplants. Cell 156, 691-704.

Garcıa-Salcedo, J. A., Perez-Morga, D., Gijon, P., Dilbeck, V., Pays, E. andNolan, D. P. (2004). A differential role for actin during the life cycle ofTrypanosoma brucei. EMBO J. 23, 780-789.

Grunfelder, C. G., Engstler, M., Weise, F., Schwarz, H., Stierhof, Y. D., Morgan,G. W., Field, M. C. and Overath, P. (2003). Endocytosis of aglycosylphosphatidylinositol-anchored protein via clathrin-coated vesicles,sorting by default in endosomes, and exocytosis via RAB11-positive carriers.Mol. Biol. Cell. 14, 2029-2040.

Guy, L. and Ettema, T. J. G. (2011). The archaeal “TACK” superphylum and theorigin of eukaryotes. Trends Microbiol. 19, 580-587.

1391

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience

Hanks, S. K. and Hunter, T. (1995). Protein kinases 6. The eukaryotic proteinkinase superfamily: kinase (catalytic) domain structure and classification. FASEBJ. 9, 576-596.

Hirst, J., Schlacht, A., Norcott, J. P., Traynor, D., Bloomfield, G., Antrobus, R.,Kay, R. R., Dacks, J. B. and Robinson, M. S. (2014). Characterization of TSET,an ancient and widespread membrane trafficking complex. Elife 3, e02866.

Hirumi, H. and Hirumi, K. (1989). Continuous cultivation of Trypanosoma bruceiblood stream forms in a medium containing a low concentration of serum proteinwithout feeder cell layers. J. Parasitol. 75, 985-989.

Ishihama, Y., Oda, Y., Tabata, T., Sato, T., Nagasu, T., Rappsilber, J. and Mann,M. (2005). Exponentially modified protein abundance index (emPAI) forestimation of absolute protein amount in proteomics by the number ofsequenced peptides per protein. Mol. Cell. Proteomics 4, 1265-1272.

Jackson, A. P., Otto, T. D., Aslett, M., Armstrong, S. D., Bringaud, F., Schlacht,A., Hartley, C., Sanders, M., Wastling, J. M., Dacks, J. B. et al. (2015).Kinetoplastid phylogenomics reveals the evolutionary innovations associated withthe origins of parasitism. Curr. Biol. 26, 161-172.

Kalb, L. C., Frederico, Y. C. A., Boehm, C., Moreira, C., Soares, M. J. and Field,M. C. (2016). Conservation and divergence within the clathrin interactome ofTrypanosoma cruzi. Nature Sci. Reps. 6, 31212.

Kirchhausen, T., Owen, D. and Harrison, S. C. (2014). Molecular structure,function, and dynamics of clathrin-mediated membrane traffic. Cold Spring Harb.Perspect. Biol. 6, a016725.

Koumandou, V. L., Dacks, J. B., Coulson, R. M. R. and Field, M. C. (2007).Control systems for membrane fusion in the ancestral eukaryote; evolution oftethering complexes and SM proteins. BMC Evol. Biol. 7, 29.

Lingnau, A., Zufferey, R., Lingnau, M. and Russell, D. G. (1999). Characterizationof tGLP-1, a Golgi and lysosome-associated, transmembrane glycoprotein ofAfrican trypanosomes. J. Cell Sci. 112, 3061-3070.

Manna, P. T. and Field, M. C. (2016). Phosphoinositides, kinases and adaptorscoordinating endocytosis in Trypanosoma brucei. Commun. Integr. Biol. 8,e1082691.

Manna, P. T., Kelly, S. and Field, M. C. (2013). Adaptin evolution in kinetoplastidsand emergence of the variant surface glycoprotein coat in Africantrypanosomatids. Mol. Phylogenet. Evol. 67, 123-128.

Manna, P. T., Boehm, C., Leung, K. F., Natesan, S. K. and Field, M. C. (2014). Lifeand times: synthesis, trafficking, and evolution of VSG. Trends Parasitol. 30,251-258.

Manna, P. T., Gadelha, C., Puttick, A. E. and Field, M. C. (2015). ENTH and ANTHdomain proteins participate in AP2-independent clathrin-mediated endocytosis.J. Cell Sci. 128, 2130-2142.

Martin, W. F., Garg, S. and Zimorski, V. (2015). Endosymbiotic theories foreukaryote origin. Philos. Trans. R. Soc. B Biol. Sci. 370, 20140330.

Miller, S. E., Mathiasen, S., Bright, N. A., Pierre, F., Kelly, B. T., Kladt, N.,Schauss, A., Merrifield, C. J., Stamou, D., Honing, S. et al. (2015). CALMregulates clathrin-coated vesicle size and maturation by directly sensing anddriving membrane curvature. Dev. Cell 33, 163-175.

Morgan, G. W., Allen, C. L., Jeffries, T. R., Hollinshead, M. and Field, M. C.(2001). Developmental and morphological regulation of clathrin-mediatedendocytosis in Trypanosoma brucei. J. Cell Sci. 114, 2605-2615.

Morgan, G. W., Denny, P. W., Vaughan, S., Goulding, D., Jeffries, T. R., Smith,D. F., Gull, K. and Field, M. C. (2005). An evolutionarily conserved coiled-coilprotein implicated in polycystic kidney disease is involved in basal bodyduplication and flagellar biogenesis in Trypanosoma brucei. Mol. Cell. Biol. 25,3774-3783.

Morriswood, B. and Schmidt, K. (2015). A MORN repeat protein facilitates proteinentry into the flagellar pocket of trypanosoma brucei. Eukaryot. Cell 14,1081-1093.

Murungi, E., Barlow, L. D., Venkatesh, D., Adung’a, V. O., Dacks, J. B., Field,M. C. and Christoffels, A. (2014). A comparative analysis of trypanosomatidSNARE proteins. Parasitol. Int. 63, 341-348.

Natesan, S. K. A., Black, A., Matthews, K. R., Mottram, J. C. and Field, M. C.(2011). Trypanosoma brucei brucei: endocytic recycling is important for mouseinfectivity. Exp. Parasitol. 127, 777-783.

Obado, S. O., Field, M. C., Chait, B. T. and Rout, M. P. (2016a). High-efficiencyisolation of nuclear envelope protein complexes from trypanosomes. In TheNuclear Envelope (Methods in Molecular Biology) (ed. S. Shackleton, P. Collasand E.C. Schirmer), pp. 67-80. Springer.

Obado, S. O., Brillantes, M., Uryu, K., Zhang, W., Ketaren, N. E., Chait, B. T.,Field, M. C. and Rout, M. P. (2016b). Interactome mapping reveals theevolutionary history of the nuclear pore complex. PLoS Biol. 14, e1002365.

Oberholzer, M., Morand, S., Kunz, S. and Seebeck, T. (2006). A vector series forrapid PCR-mediated C-terminal in situ tagging of Trypanosoma brucei genes.Mol. Biochem. Parasitol. 145, 117-120.

Ounjai, P., Kim, K. D., Liu, H., Dong, M., Tauscher, A. N., Witkowska, H. E. andDowning, K. H. (2013). Architectural insights into a ciliary partition. Curr. Biol. 23,339-344.

Pearse, B. M. (1976). Clathrin: a unique protein associated with intracellular transferof membrane by coated vesicles. Proc. Natl. Acad. Sci. USA 73, 1255-1259.

Price, H. P., Stark, M., Smith, B. and Smith, D. F. (2007). TbARF1 influenceslysosomal function but not endocytosis in procyclic stage Trypanosoma brucei.Mol. Biochem. Parasitol. 155, 123-127.

Raymann, K., Brochier-Armanet, C. and Gribaldo, S. (2015). The two-domaintree of life is linked to a new root for the Archaea. Proc. Natl. Acad. Sci. USA 112,6670-6675.

Redmond, S., Vadivelu, J. and Field, M. C. (2003). RNAit: an automated web-based tool for the selection of RNAi targets in Trypanosoma brucei.Mol. Biochem.Parasitol. 128, 115-118.

Ricotta, D., Conner, S. D., Schmid, S. L., von Figura, K. and Honing, S. (2002).Phosphorylation of the AP2 mu subunit by AAK1 mediates high affinity binding tomembrane protein sorting signals. J. Cell Biol. 156, 791-795.

Schlacht, A., Mowbrey, K., Elias, M., Kahn, R. A. and Dacks, J. B. (2013). Ancientcomplexity, opisthokont plasticity, and discovery of the 11th subfamily of Arf GAPproteins. Traffic 14, 636-649.

Schlacht, A., Herman, E. K., Klute, M. J., Field, M. C. and Dacks, J. B. (2014).Missing pieces of an ancient puzzle: evolution of the eukaryotic membrane-trafficking system. Cold Spring Harb. Perspect. Biol. 6, a016048.

Schwede, A., Macleod, O. J. S., MacGregor, P. and Carrington, M. (2015). Howdoes the VSG coat of bloodstream form african trypanosomes interact withexternal proteins? PLoS Pathog. 11, e1005259.

Sheader, K., Vaughan, S., Minchin, J., Hughes, K., Gull, K. and Rudenko, G.(2005). Variant surface glycoprotein RNA interference triggers a precytokinesiscell cycle arrest in African trypanosomes. Proc. Natl. Acad. Sci. USA 102,8716-8721.

Shimogawa, M. M., Saada, E. A., Vashisht, A. A., Barshop,W. D., Wohlschlegel,J. A. and Hill, K. L. (2015). Cell surface proteomics provides insight into stage-specific remodeling of the host-parasite interface in Trypanosoma brucei. Mol.Cell. Proteomics 14, 1977-1988.

Spitznagel, D., O’Rourke, J. F., Leddy, N., Hanrahan, O. and Nolan, D. P. (2010).Identification and characterization of an unusual class I myosin involved in vesicletraffic in Trypanosoma brucei. PLoS ONE 5, e12282.

Stachowiak, J. C., Schmid, E. M., Ryan, C. J., Ann, H. S., Sasaki, D. Y., Sherman,M. B., Geissler, P. L., Fletcher, D. A. and Hayden, C. C. (2012). Membranebending by protein-protein crowding. Nat. Cell Biol. 14, 944-949.

Steverding, D. (2000). The transferrin receptor of Trypanosoma brucei. Parasitol.Int. 48, 191-198.

Sunter, J., Wickstead, B., Gull, K., Carrington, M., Wirtz, E., Clayton, C., Wirtz,E., Leal, S., Ochatt, C., Cross, G. et al. (2012). A new generation of T7 RNApolymerase-independent inducible expression plasmids for Trypanosoma brucei.PLoS ONE 7, e35167.

Szoor, B., Haanstra, J. R., Gualdron-Lopez, M. and Michels, P. A. M. (2014).Evolution, dynamics and specialized functions of glycosomes in metabolism anddevelopment of trypanosomatids. Curr. Opin. Microbiol. 22, 79-87.

Tanabe, K., Torii, T., Natsume, W., Braesch-Andersen, S., Watanabe, T. andSatake, M. (2005). A novel GTPase-activating protein for ARF6 directly interactswith clathrin and regulates clathrin-dependent endocytosis. Mol. Biol. Cell. 16,1617-1628.

Tazeh, N. N., Silverman, J. S., Schwartz, K. J., Sevova, E. S., Sutterwala, S. S.and Bangs, J. D. (2009). Role of AP-1 in developmentally regulated lysosomaltrafficking in Trypanosoma brucei. Eukaryot. Cell 8, 1352-1361.

Tomavo, S. (2014). Evolutionary repurposing of endosomal systems for apicalorganelle biogenesis in Toxoplasma gondii. Int. J. Parasitol. 44, 133-138.

Umeda, A., Meyerholz, A. and Ungewickell, E. (2000). Identification of theuniversal cofactor (auxilin 2) in clathrin coat dissociation. Eur. J. Cell Biol. 79,336-342.

Urbaniak, M. D., Guther, M. L. S. and Ferguson, M. A. (2012). Comparative SILACproteomic analysis of Trypanosoma brucei bloodstream and procyclic lifecyclestages. PLoS ONE 7, e36619.

Venkatesh, D., Boehm, C., Barlow, L. D., Nankissor, N. N., O‘Reilly, A. J., Kelly,S., Dacks, J. B., and Field, M. C. (2017). Evolution of the endomembranesystems of trypanosomatids: conservation and specialisation. J. Cell Sci. 8,1421-1434.

Wirtz, E., Leal, S., Ochatt, C. and Cross, G. M. (1999). A tightly regulated inducibleexpression system for conditional gene knock-outs and dominant-negativegenetics in Trypanosoma brucei. Mol. Biochem. Parasitol. 99, 89-101.

Yen,W.-L., Legakis, J. E., Nair, U. and Klionsky, D. J. (2007). Atg27 is required forautophagy-dependent cycling of Atg9. Mol. Biol. Cell 18, 581-593.

Zoltner, M., Leung, K. F., Alsford, S., Horn, D. and Field, M. C. (2015). Modulationof the surface proteome through multiple ubiquitylation pathways in AfricanTrypanosomes. PLoS Pathog. 11, e1005236.

1392

RESEARCH ARTICLE Journal of Cell Science (2017) 130, 1379-1392 doi:10.1242/jcs.191478

Journal

ofCe

llScience